Tautomeric and Conformational Properties of Benzoylacetone, CH3–C

Mar 5, 2012 - Emsley , J. In Structure and bonding; Clarke , M. J. ; J. B.G. ; Ibers , J. A. ; Jorgensen , C. K. ; Mingos , D. M. P. ; Neilands , J. B...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/JPCA

Tautomeric and Conformational Properties of Benzoylacetone, CH3− C(O)−CH2−C(O)−C6H5: Gas-Phase Electron Diffraction and Quantum Chemical Study Natalya V. Belova,*,† Georgiy V. Girichev,† Heinz Oberhammer,‡ Trang Nguen Hoang,† and Sergey A. Shlykov† †

Ivanovo State University of Chemistry and Technology, Engelsa av., 7, Ivanovo 153460, Russia Institut für Physikalische und Theoretische Chemie, Universität Tübingen, Auf der Morgenstelle 8, 72076 Tübingen, Germany



S Supporting Information *

ABSTRACT: Tautomeric and structural properties of benzoylacetone, CH3−C(O)−CH2−C(O)−C6H5, have been studied by gas-phase electron diffraction (GED) and quantum chemical calculations (B3LYP and MP2 approximation with different basis sets up to aug-cc-pVTZ). Analysis of GED intensities resulted in the presence of 100% enol tautomer at 331(5) K. The existence of two possible enol conformers in about equal amounts is confirmed by both GED and quantum chemical results. In both conformers the enol ring possesses Cs symmetry with a strongly asymmetric hydrogen bond. The experimental geometric parameters are reproduced very closely by the B3LYP/cc-pVTZ method.



INTRODUCTION Over many years, β-diketones (β-dicarbonyl compounds) of type R1C(O)CH2C(O)R2 have been of considerable interest because they are important organic reagents and possess interesting properties.1−4 Special attention has been paid to the keto−enol tautomerism of β-diketones, the structural properties of both keto and enol forms, and the nature of the strong OH···O hydrogen bond in the enol form. It is well-known that the preference of the enol or keto tautomer of β-diketones depends strongly on the temperature, solvent and phase state.3 Furthermore, the keto/enol ratio also depends strongly on the substituents R1 and R2. In this work we report tautomeric, conformational, and structural properties of benzoylacetone (BA), which possesses two nonequivalent substituents R1 = CH3 and R2 = C6H5. For the β-diketones with R1 ≠ R2 two different enol forms can occur, with O−H bond adjacent to R1 or to R2. Furthermore, different conformations are feasible for the diketo tautomer, depending on the relative orientation of the CO bonds (Scheme 1). Several NMR studies of BA5−11 show the predominance of the enol tautomer in solutions from 77% of enol form in DMSO8 to 100% of enol in CCl4.10 Moreover, all authors point out that the presence of a phenyl ring as substituent increases the enol content when going from acetylacetone (AcAc) to benzoylacetone and dibenzoylmethane (DBM). Whereas Burgett et al.10 suppose that the stabilization of the enol tautomer in the case of BA and DBM is the result of inductive effects of the phenyl rings, Emsley3 assumes that steric reasons, depending on the nature of the β-substituents, may be © 2012 American Chemical Society

responsible for the varying percentage of the enol form. However, both relationships are not very clear. Near vacuum UV absorption spectra of BA12 have been measured in heptane, in acetonitrile, and in ethanol at room temperature. The analysis of spectra clearly shows that BA exists in the enol form with rather strong π-electron interaction between the enol ring and the phenyl ring. A spectrophotometric study of tautomeric properties of BA in aqueous solution13 (with ionic strength 0.1) at 25 °C resulted in an equilibrium constant of enolization K = [enol]/ [keto] of only 0.72, which implies a rather equal content of both keto and enol tautomeric forms in aqueous solution. In contrast, K = [enol]/[keto] = 6 for DBM at the same conditions.13 One of the widely discussed structural problems is the hydrogen bonding of the enol forms of β-diketones and the related intramolecular enol−enol tautomerism. The key controversy has been the shape of the O−H...O potential function, either a single-minimum potential, also described as a “symmetrical”, strongly delocalized system, or a doubleminimum potential, corresponding to two tautomeric forms. Furthermore, Chan et el.14 suggested equilibrium between a symmetrical and normal tautomeric system. According to the 17O NMR studies of BA in solutions,15 59% of enolization occurs at the phenyl side. This result suggests the Received: November 14, 2011 Revised: March 5, 2012 Published: March 5, 2012 3428

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

Scheme 1. Enol Tautomers (Above) and Possible Conformers of Diketo Tautomer of R1C(O)−CH2−C(O)R2 Compounds (Below)a

“s” stands for synperiplanar, sp, (which corresponds to dihedral angles τ(OCCC) of 0 ± 30°) or synclinal, sc, (τ(OCC C) = 60 ± 30°) and “a” for anticlinal, ac, (τ(OCCC) = 120 ± 30°) or antiperiplanar, ap, ((τ(OCCC) = 180 ± 30°).

a

coexistence of two enol conformers (Figure 1) with the O−H bond adjacent to methyl (enol1) or adjancent to phenyl (enol2) and both forms occur in comparable amounts. Variable-temperature NMR study of the enol forms of BA16 also showed that a double-minimum exists and the tautomeric system is favored until 181 K. This result is in a agreement with conclusions based on the resonance-assisted hydrogen bond theory, introduced by Gilli et al.17 The equilibrium constant of enol2 ↔ enol1 (Scheme 1) tautomerization of BA in CH2Cl2 at 227 K was found to be K = 0.89, which implies almost equal amounts of both enol forms (53% of enol2). On the other hand, based on NMR studies of BA in CDCl3 Matsuzawa et al.9 concluded that the enolic proton is preferentially located near the oxygen at the side of the phenyl group. However, in the solid state the enol structure of BA with an almost symmetrical enol ring was found.18,19 According to Xray and neutron diffraction results18,19 the two C−C and the two C−O bonds of the enol ring do not differ significantly in length and the enol hydrogen is located in a broad singleminimum potential well. The authors19 conclude that the symmetry of the enol ring arises from a single π-delocalized structure rather than from the superposition of two localized structures. In contrast to these results, the other X-ray study of BA20 resulted in an enol structure with enolic hydrogen located in an asymmetric position, adjacent to a phenyl ring. The two C−C and the two C−O bonds are significantly different in length. Both phenyl and enol rings of BA in the crystal are planar and are twisted relative to each other by an angle of 6.3° 20 or 6.4°.19 A quantum chemical investigation (DFT(B3LYP) with different basis sets)21 was performed only for the enol conformers of benzoylacetone. IR and Raman spectra of BA (solid and solution in CCl4) and its deuterated analogue were clearly assigned to the enol tautomer. Natural bond orbital analysis (NBO) was used to compare some properties of BA with dibenzoylmethane (DBM) and acetylacetone (AcAc), such as molecular stability and the hydrogen bond strength.21

Figure 1. Molecular structure of two enol (above) and (ac, ac) diketo (below) forms of benzoylacetone.

The following trend in hydrogen bond strength has been obtained: AcAc < BA < DBM based on vibrational spectroscopy and quantum chemical results.21 Tayyari et al.21 suppose that the increase of the hydrogen bond strength in this series is the result of enol ring stabilization through resonance with phenyl groups.



QUANTUM CHEMICAL CALCULATIONS All quantum chemical calculations were performed with the GAUSSIAN03 program set.22 To detect all possible diketo conformers of BA, the potential energy surface has been scanned at the HF/6-31G(d,p) level. The torsional angles τ(O1C2C1C3) and τ(O2C3C1C2) were varied in steps of 20° with full optimization of all other parameters (see Figure 1 for atom numbering). This surface possesses three minima corresponding to (ac, ac), (ac, sp), and (sp, ac) conformers (Scheme 1). The geometries of these diketo conformers and of two possible enol conformers were fully optimized with the B3LYP and MP2 methods with 6-31G(d,p) and cc-pVTZ basis sets. The relative energies (ΔE = Eketo − Eenol) and relative free energies (ΔG0 = G0keto − G0enol) obtained with the different computational methods are summarized in Table 1 along with 3429

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

Table 2 together with experimental results. Vibrational amplitudes and corrections, Δr = rh1 − ra, were derived from theoretical force fields (B3LYP/cc-pVTZ) by Sipachev’s method (approximation with nonlinear transformation of Cartesian coordinates into internal coordinates), using the program SHRINK.23−25 Relevant values for the enol forms are listed in Table 3 and a full list of interatomic distances, vibrational amplitudes, and vibrational corrections for the enol forms (excluding nonbonded distances involving hydrogen) are available as Supporting Information. The NBO 5G program,26 implemented for natural orbital analysis in PC GAMESS,27 was used to obtain the net atomic charges, and to study the effect of hyperconjugation on the structure of the two enol conformers and (ac, ac) diketo conformer. B3LYP/aug-cc-pVTZ wave functions were used in the NBO analyses. Relevant values of second order interaction energies (E(2)) between donor−acceptor orbitals are collected in Table 4.

Table 1. Torsion Angles, Optimized Relative Energies, and Gibbs Free Energies of the Enol Tautomer and Diketo Conformers of Benzoylacetonea enol1b τ(O1C2C1C3) τ(O2C3C1C2) τ(O1C2C4C6) τ(O2C3C5H8) Erel, kcal/mol Grel0, kcal/mol

0.0 0.0 −24.0 179.8 0.0 0.0

τ1 (O1C2C1C3) τ2 (O2C3C1C2) τ(O1C2C4C6) τ(O2C3C5H8) Erel, kcal/mol Grel0, kcal/mol

0.0 0.0 −12.7 180.0 0.0 0.0

τ1 (O1C2C1C3) τ2 (O2C3C1C2) τ(O1C2C4C6) τ(O2C3C5H8) Erel, kcal/mol Grel0, kcal/mol

0.0 0.0 −8.6 179.8 0.0 0.0

τ1 (O1C2C1C3) τ2 (O2C3C1C2) τ(O1C2C4C6) τ(O2C3C5H8) Erel, kcal/mol Grel0, kcal/mol

0.0 0.0 −8.6 180.0 0.0 0.0

enol2b

diketo (sp, ac)

MP2/6-31G(d,p) 0.0 2.9 0.0 105.5 −28.7 0.1 3.0 2.0 0.50 2.23 −0.16 0.05 MP2/cc-pVTZ 0.0 2.5 0.0 102.7 −16.8 0.2 −0.1 −5.9 0.47 6.76 −0.19 4.58 B3LYP/6-31G(d,p) 0.0 7.7 0.0 112.2 −12.7 −1.0 20.4 5.0 0.64 7.03 −0.50 5.50 B3LYP/cc-pVTZ 0.0 8.4 0.0 108.3 −11.9 −1.3 24.6 1.0 0.57 7.53 −0.62 5.65

diketo (ac,ac)

diketo (ac, sp)

90.5 101.6 15.2 −12.1 0.45 −0.51

107.7 −8.7 −0.4 −4.6 2.67 0.45

92.3 102.7 11.3 −13.1 5.03 4.07

106.9 −10.6 0.8 −6.6 7.11 4.89

91.9 100.5 11.8 −11.6 5.55 4.93

c

89.1 100.1 11.5 −10.9 6.27 5.32

c



STRUCTURE ANALYSIS The heaviest ion in the mass spectrum was the parent ion [C10H10O2]+ (Table 7). This proves that monomers are present in the vapor at the conditions of the GED experiment. The experimental radial distribution function (Figure 2) was derived by Fourier transformation of the experimental intensities. Figure 2 demonstrates that the theoretical radial distribution functions for enol and keto tautomers of benzoylacetone, CH3−C(O)−CH2−C(O)−C6H5 differ strongly in the region of long nonbonded distances. The experimental curve can be reproduced reasonably well only with the enol conformers. Therefore, in the least-squares analysis only the enol forms were considered. LS analysis was carried out for each enol conformer separately. The theoretical sM(s) functions were calculated with the following assumptions. Independent rh1 parameters were used to describe the molecular structure. Starting parameters from B3LYP/cc-pVTZ calculation were refined by a least-squares procedure of the molecular intensities. The differences between all C−H and O−H bond lengths, between all C−C bond lengths were constrained to calculated values (B3LYP/cc-pVTZ). A planar skeleton with Cs symmetry of the enol ring was assumed. Vibrational amplitudes were refined in groups with fixed differences. With the abovementioned assumptions four bond distances and seven bond angles (Table 2) were refined simultaneously with ten groups of vibrational amplitudes (Table 3). The following correlation coefficients had values larger than |0.8|: r(C1C2)/r(C2O1) = 0.81, r(C1C2)/r(C3O2) = −0.82, r(C2O1)/r(C3O2) = −0.92, ∠(C2C1C3)/∠(C1C2O1) = −0.90, ∠(O1C2C4)/ ∠(O2C3C5) = −0.88 for the enol1 conformer and r(C2O1)/r(C3O2) = −0.93, ∠(C2C1C3)/∠(C1C2O1) = −0.80, ∠(O1C2C4)/∠(O2C3C5) = −0.86 for the enol2. The best agreement factor in the least-squares analysis resulted for the enol1 form with Rf = 3.2% and for the enol2 form with Rf = 3.0%. Final results of the least-squares analysis for both enol conformers are given in Table 2 (geometric parameters) and Table 3 (vibrational amplitudes). The refined geometric parameters of the two enol forms are rather similar to those predicted by quantum chemical calculations. Moreover, the analogous structural parameters are very close in both conformers. Thus, it is not possible to choose the preferred enol conformer either by quantum chemical calculations or by GED results.

a

All values of angles in deg. bSee Figure 1. cNot a stable conformer according to this method.

the skeletal OCCC dihedral angles and the torsional angles of the methyl group and phenyl ring. All geometric parameters obtained with these methods are very close, except for the results of the MP2/6-31G(d,p) method for the torsional angles of the phenyl groups in the enol tautomer and the OCCC angles in the (sp, ac) diketo conformer. The values in Table 1 demonstrate that predicted relative energies of diketo and enol tautomers depend strongly on the computational method. For comparison of calculated and experimental tautomeric composition the Gibbs free energies must be applied, instead of the relative energies. G0 values include zero point energies, temperature corrections, and entropies. Predictions concerning the tautomeric equilibrium depend strongly on the computational method, as well. Whereas both B3LYP methods and the MP2/cc-pVTZ calculation predict a strong preference of the enol tautomer, the MP2 method with small basis sets predicts Grel0 values (+0.05, −0.51, and +0.45 kcal/mol for (sp, ac), (ac, ac), and (ac, sp) conformers, respectively), which correspond to preference of the diketo tautomer. All methods predict a rather equal content of the two possible enol conformers. Although the calculated relative energy of enol1 is slightly lower at all levels of theory, Grel0 values predict a slight preference of enol2 with enolic hydrogen adjacent to the phenyl group. The geometric parameters for two enol conformers which were derived with the B3LYP/cc-pVTZ method are listed in 3430

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

Table 2. Experimental and Calculated Geometric Parameters of the Enol Tautomers of Benzoylacetonea enol1 r(C1−C2) r(C1−C3) r(C9−C10) r(C6−C7) r(C7−C8) r(C8−C9) r(C4−C6) r(C4−C10) r(C2−C4) r(C3−C5) r(C3−O2) r(C2−O1) r(C1−H2) r(C6−H3) r(C10−H7) r(C9−H6) r(C7−H4) r(C8−H5) r(C5−H8) r(C5−H9) r(C5−H10) r(O1−H1) r(O2−H1) r(O1···O2) ∠C2C1C3 ∠C1C2O1 ∠C1C3O2 ∠O1C2C4 ∠O2C3C5 ∠C2O1H1 ∠C3O2H1 ∠C2C4C6 ∠C6C4C10 ∠C7C6C4 ∠C8C7C6 ∠C9C8C7 ∠C10C9C8 ∠C4C10C9 ∠C3C5H8 ∠C3C5H9 ∠C3C5H10 C4C2C1O1 C5C3C1O2 C6C4C2O1 H8C5C3O2

enol2

GED (rh1, ∠h1)b

B3LYP/cc-pVTZ

GED (rh1, ∠h1)b

B3LYP/cc-pVTZ

X-ray (ref 20)

X-ray (ref 19)

neutron diffraction (ref 19)

1.443(3) p1c 1.373(3) (p1) 1.392(3) (p1) 1.390(3) (p1) 1.396(3) (p1) 1.394(3) (p1) 1.403(3) (p1) 1.402(3) (p1) 1.498(3) (p1) 1.496(3) (p1) 1.308(3) p2 1.256(3) p3 1.080(4) p4 1.084(4) (p4) 1.084(4) (p4) 1.086(4) (p4) 1.086(4) (p4) 1.086(4) (p4) 1.091(4) (p4) 1.096(4) (p4) 1.096(4) (p4) 1.574(4)e 1.014(4) (p4) 2.507(5)e 120.1(8) p5 120.7(8) p6 122.1(8) p7 119.6(9) p8 116.6(9) p9 101.1e 105.3d 118.2(12) p10 118.8d 120.6d 120.1d 119.8d 120.1d 120.6d 111.7d 109.7d 109.7d 179.5d 179.9d 0.2(8.0) p11 180.0d

1.439 1.369 1.388 1.386 1.392 1.390 1.399 1.398 1.494 1.492 1.321 1.251 1.076 1.080 1.080 1.082 1.082 1.082 1.087 1.091 1.091 1.585 1.009 2.510 120.4 120.6 122.1 118.5 114.0 101.6 105.3 118.1 118.8 120.6 120.1 119.8 120.1 120.6 111.7 109.7 109.7 179.5 179.9 8.6 180.0

1.378(3) p1 1.439(3) (p1) 1.390(3) (p1) 1.391(3) (p1) 1.393(3) (p1) 1.395(3) (p1) 1.403(3) (p1) 1.403(3) (p1) 1.479(3) (p1) 1.512(3) (p1) 1.245(3) p2 1.313(3) p3 1.083(4) p4 1.086(4) (p4) 1.086(4) (p4) 1.088(4) (p4) 1.088(4) (p4) 1.088(4) (p4) 1.094(4) (p4) 1.099(4) (p4) 1.099(4) (p4) 1.018(4) (p4) 1.578(4)e 2.520(5)e 120.9(7) p5 120.4(7) p6 121.8(7) p7 115.7(9) p8 120.6(9) p9 105.7d 99.7e 119.4(13) p10 118.7d 120.5d 120.2d 119.7d 120.2d 120.6d 109.7d 109.0d 109.0d 179.2d 179.0d 11.8(5.3) p11 24.6d

1.375 1.435 1.386 1.387 1.390 1.391 1.399 1.400 1.475 1.509 1.247 1.324 1.077 1.079 1.080 1.082 1.082 1.082 1.087 1.092 1.092 1.012 1.573 2.507 120.6 120.5 121.9 114.7 119.1 105.7 100.0 119.4 118.7 120.5 120.2 119.7 120.2 120.6 109.7 109.0 109.0 179.2 179.0 11.9 24.6

1.383 1.405 1.376 1.388 1.380 1.379 1.396 1.401 1.487 1.520 1.293 1.312 0.96 0.90 0.94 0.91 0.95 0.94

1.402(2) 1.408(2) 1.391(1) 1.392(1) 1.395(2) 1.395(2) 1.405(1) 1.402(1) 1.481(1) 1.495(2) 1.286(2) 1.293(2) 1.076 1.081 1.079 1.087 1.092 1.086 1.060 1.064 1.050 1.247 1.325 2.499(2) 119.8(2) 121.2(2) 122.0(2) 116.2(2) 117.3(2) 102.6 101.1 118.9(1) 119.4(2) 120.1(1) 120.3(1) 119.7(2) 120.5(1) 120.0(1) 113.0 110.7 109.8

1.405(4) 1.414(4) 1.402(4) 1.394(4) 1.404(4) 1.387(4) 1.404(3) 1.406(3) 1.483(4) 1.499(4) 1.286(4) 1.293(4) 1.076(6) 1.085(6) 1.079(6) 1.090(6) 1.090(6) 1.091(7) 1.064(7) 1.062(8) 1.051(6) 1.245(11) 1.329(11) 2.502(4) 119.7(2) 120.9(3) 122.1(3) 116.4(3) 117.0(3) 103.2(4) 101.2(4) 119.0(2) 119.5(3) 120.3(2) 119.9(3) 120.1(3) 120.4(3) 119.8(3) 112.6(5) 110.8(4) 109.8(4)

1.18 1.40 2.502 121.9 120.4 120.6 115.6 117.5 104.0 119.8 118.4 120.2 120.2 119.9 120.6 120.5

a Distances in Å and angles in degrees. For atom numbering, see Figure 1. bUncertainties in rh1 σ = (σsc2 + (2.5σLS)2)1/2 (σsc = 0.002r, σLS = standard deviation in least-squares refinement), for angles σ = 3σLS. cpi = parameter refined independently. (pi) = parameters calculated from the independent parameter pi by a difference Δ = pi − (pi) from the quantum chemical calculations dNot refined. eDependent parameter.

As mentioned above, different experimental and theoretical studies for BA lead to controversial results concerning the symmetry of the hydrogen bond. Some studies result in an asymmetric O−H...O hydrogen bond with Cs symmetry of the enol ring,9,15−17 other studies in a symmetric O···H···O bond with near C2v symmetry of the enol skeleton.18,19 Therefore, a model with symmetric hydrogen bond and C2v symmetry of the enol skeleton was tested in the GED analysis. This refinement lead to Rf = 4.8%, considerably higher than Rf = 3.0% and 3.2%, obtained for the models with asymmetric hydrogen bond. Thus,

the GED experiment confirms our quantum chemical calculations, which predict an asymmetric enol structure with a localized O−H bond. Calculated potential functions for the hydrogen position between the two oxygen atoms are discussed below.



DISCUSSION The GED experiment for benzoylacetone, CH3−C(O)−CH2− C(O)−C6H5 is consistent with the presence of 100% enol form in the gas phase at 331(5) K. This result is in agreement with 3431

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

Table 4. Relevant Second-Order Perturbation Energies E(2) (Donor−Acceptor), kcal/mol

Table 3. Interatomic Distances, Vibrational Amplitudes, and Vibrational Corrections for the Enol Tautomera,b rh1, Å O2−H1 C1−H2 C6−H3 C2−O1 C3−O2 C1−C3 C4−C10 C1−C2 C3−C5 C2−C4 C1−O2 C1−O1 C4−O1 C5−O2 C2−C3 C1−C5 O1−O2 C3−O1 C2−O2 C2−C5 C5−O1 C4−O2 C4−C5

1.014(4) 1.080(4) 1.084(4) 1.256(3) 1.308(3) 1.373(3) 1.402(3) 1.443(3) 1.496(3) 1.498(3) 2.347(5) 2.347(5) 2.382(5) 2.387(11) 2.441(5) 2.501(10) 2.507(5) 2.775(6) 2.797(6) 3.827(10) 4.271(9) 4.293(9) 5.045(14)

O1−H1 C1−H2 C6−H3 C3−O2 C2−O1 C1−C2 C4−C10 C1−C3 C2−C4 C3−C5 C1−O1 C1−O2 C4−O1 C5−O2 C2−C3 O1−O2 C1−C5 C2−O2 C3−O1 C2−C5 C4−O2 C5−O1 C4−C5

1.018(4) 1.083(4) 1.086(4) 1.245(3) 1.313(3) 1.378(3) 1.403(3) 1.439(3) 1.479(3) 1.512(3) 2.336(5) 2.347(5) 2.365(5) 2.399(10) 2.450(5) 2.520(5) 2.525(11) 2.806(6) 2.783(6) 3.804(10) 4.284(9) 4.291(9) 5.045(14)

l (GED) Enol1 0.080(2) l1 0.075(2) l1 0.075(2) l1 0.040(2) l1 0.044(2) l1 0.044(2) l2 0.045(2) l2 0.048(2) l2 0.049(2) l2 0.050(2) l2 0.059(4) l3 0.060(4) l3 0.062(4) l3 0.068(4) l3 0.061(4) l3 0.069(4) l3 0.101(4) l3 0.074(4) l4 0.076(4) l4 0.075(6) l5 0.086(7) l6 0.089(7) l6 0.076(14) l7 Enol2 0.083(2) l1 0.074(2) l1 0.075(2) l1 0.041(2) l1 0.045(2) l1 0.046(2) l2 0.045(2) l2 0.050(2) l2 0.048(2) l2 0.051(2) l2 0.055(3) l3 0.055(3) l3 0.059(3) l3 0.074(3) l3 0.058(3) l3 0.114(3) l3 0.079(3) l3 0.080(3) l4 0.084(3) l4 0.078(5) l5 0.083(7) l6 0.087(7) l6 0.097(13) l7

l (B3LYP/ccpVTZ)

rh1 − ra

0.081 0.075 0.075 0.041 0.044 0.044 0.046 0.048 0.050 0.050 0.055 0.055 0.058 0.064 0.056 0.064 0.096 0.074 0.076 0.066 0.080 0.082 0.083

−0.0008 0.0018 0.0017 −0.0004 0.0008 −0.0002 0.0004 0.0019 0.0002 0.0010 0.0010 0.0041 0.0039 0.0063 0.0036 0.0018 −0.0038 0.0022 0.0021 0.0093 0.0092 0.0122 0.0185

0.084 0.075 0.075 0.042 0.046 0.046 0.046 0.050 0.049 0.051 0.055 0.055 0.059 0.074 0.058 0.114 0.079 0.080 0.085 0.073 0.085 0.090 0.099

−0.0035 0.0024 0.0016 −0.0019 0.0025 −0.0016 0.0004 0.0039 0.0007 0.0009 0.0037 0.0035 0.0043 0.0134 0.0060 0.0069 −0.0008 0.0069 0.0083 0.0108 0.0137 0.0221 0.0163

donor

E(2)

acceptor Enol form enol1

π(C1−C2) π(C1−C3) π(C4−C6) π(C4−C10) Lp(2) O1 Lp(2) O2 Lp(2) O2 Lp(2) O2 Lp(2) O1 Lp(2) O1 π(C4−C10) Lp(2) O1 Lp(2) O1 Lp(2) O2 Lp(2) O2

π*(C3−O2) π*(C2−O1) π*(C1−C2) π*(C2−O1) π*(C1−C2) σ*(C3−C5) σ*(O1−H1) π*(C1−C3) σ*(C2−C4) σ*(O2−H1) Keto (ac, ac) form π*(C2−O1) σ*(C1−C2) σ*(C2−C4) σ*(C1−C3) σ*(C3−C5)

enol2 34.31

32.99 20.02 22.28 48.63 17.91 34.43 50.37 17.93 31.67 20.01 21.64 19.72 23.01 20.46

Figure 2. Experimental (dots) and calculated radial distribution functions and difference curve for the enol2.

we can suppose the existence of two enol conformers in about equal amounts. Both conformers possess an asymmetric skeleton with localized O−H bond and strong intramolecular O−H···O hydrogen bond. GED structures are in perfect agreement with quantum chemical predictions. Table 2 compares gas-phase and crystal structures of BA. Whereas the enol ring in the gas phase clearly possesses an asymmetric O−H···O hydrogen bond, the differences between the two C−C bond lengths as well as between the two C−O bond distances in the crystal are smaller. Although the X-ray and neutron diffraction studies by Herbstein et al.19 result in the position of enolic H-atom close to the middle of the O···O distance, according to the X-ray study by Semmingsen20 the enol ring is asymmetric with an asymmetric hydrogen bond. On the other hand, a combined NMR study of the enol form of several compounds16 results in a two-well potential for BA, as well as for AcAc and DBM in the liquid phase. In an attempt to explain the difference between NMR results for solutions and X-ray results, Borisov et al.16 suggest that a reduction of the twist angle in the crystal leads to more steric compression and to a shorter O···O distance that ultimately should lead to a longer O−H distance and to a more symmetrical structure. Figure 3 presents calculated potential curves for the position of

a

Values in Å. Error limits for the amplitudes are 3σ values. For atom numbering see Figure 1. bFull table of interatomic distances, vibrational amplitudes and vibrational corrections for the enol form (excluding nonbonded distances involving hydrogen) is available as Supporting Information

the predictions by most quantum chemical calculations. Only the MP2/6-31G(d,p) method predicts the preference of the diketo form. As mentioned in previous investigations,4,28 this level of theory fails in predicting the tautomeric composition of β-diketones. According to GED and quantum chemical results 3432

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

distribution in the enol ring which is evidence of a weak electron donating character of the phenyl ring. Table 4 summarizes the relevant values of second-order interaction energies (E(2)) between donor−acceptor orbitals in both enol conformers and (ac, ac) diketo form of BA obtained in the present study. The E(2) values show that hyperconjugation between the phenyl group and π* orbitals of the adjacent double bonds exists not only in the enol but also in diketo form. The energies of such interactions are rather similar in all three cases. Thus, these hyperconjugations cannot be considered to be an additional stabilization of the enol form. Obviously, interactions of the phenyl groups with the double bonds just keep the phenyl rings nearly coplanar with the enol skeleton, despite the steric repulsions mentioned above. Table 5 summarizes the calculated values that characterize the hydrogen bond in the enol form of AcAc, BA, and DBM. In

Figure 3. Calculated (B3LYP/6-31G(d,p)) potential curve for the hydrogen atom between the two oxygen atoms. The position of the hydrogen atom is described by the difference between O1···H and O2···H distances.

Table 5. Values Characterizing Hydrogen Bond in the Enol Form of β-Diketonesa

the hydrogen atom between two oxygen atoms derived by B3LYP calculations with 6-31G(d,p) basis sets. The two minima occur for localized hydrogen bonds with r(O1−H) = 1.013 Ǻ and r(O2−H) = 1.012 Ǻ , the maximum for r(O1···H) = r(O2···H) = 1.219 Ǻ . The height of this barrier leads to a predominantly localized O−H bond in the GED experiment. On the other hand, the barriers of 2.68 and 2.00 kcal/mol are low enough to allow exchange of the hydrogen position during the X-ray diffraction experiment at room temperature, leading to an intermediate position of the hydrogen atom. Furthermore, the properties of the O−H···O group in the crystal may differ from those of the free molecule due to intermolecular hydrogen bonds between neighboring molecules. According to the IR study the value of ν(O−H) = 2650 cm−1 (in CCl4)21,29 suggests the presence of strong intramolecular hydrogen bond with predominantly localized position of the enolic H-atom. It should be noted that both enol structures deviate slightly from Cs symmetry due the rotation of the phenyl ring relative to the enol skeleton, although hyperconjugation tends to keep the phenyl rings coplanar with the enol skeleton. According to B3LYP/cc-pVTZ calculations the torsional angles τ(C6C4C2O1) are similar in the enol1 and enol2 (8.6° and 11.9°). The GED experiment results in a larger difference (0.2(8.0)° and 11.8(5.3°)), but within the experimental uncertainties both angles are again similar. Such a rotation of the phenyl ring occurs also in the crystal (τ = 6.3−6.4°). This deviation from Cs symmetry is due to steric repulsion between enol and phenyl carbon and hydrogen atoms. In the enol2 form the C...C distances (r(C2−C6) = 2.485 Ǻ , r(C1−C10) = 2.999 Ǻ ) are very short and H...H distance, r(H2−H7) = 2.037 Ǻ is definitely shorter than the sum of Van-der-Waals radii (2.40 Å). The calculated potential curve for the rotation of phenyl group shows that the maximum of energy corresponds to the rotational angle 90°. The height of the barrier calculated with B3LYP/cc-pVTZ is 4.65 kcal/mol. Tayyari et al.21 suggest that the phenyl group increases the H-bond strength through conjugation with the enol ring resulting in the following trend in hydrogen bond strength: AcAc < BA < DBM. Furthermore, the authors21 have concluded that, in the enol2 conformer of BA, the phenyl ring is conjugated to the CCCO fragment whereas in the enol1 the phenyl group is conjugated only to the CO bond. On the basis of an NBO analysis of the enol forms, Tayyari et al.21,30 note that the substitution of CH3 groups in AcAc by phenyl in BA and DBM slightly modifies the charge

benzoylacetone r(O···O) r(O−H) r(O···H) Q(O−H) Q(O···H) q(O1) q(O2) q(R1) q(R2) q(ring)b ω(O−H)

acetylacetone

enol1

enol2

dibenzoylmethane

2.533 1.006 1.615 0.615 0.111 −0.655 −0.635 0.035 0.008 −0.257 2990

2.511 1.010 1.588 0.608 0.119 −0.643 −0.652 0.023 0.034 −0.269 2934

2.506 1.013 1.571 0.600 0.127 −0.653 −0.639 0.073 0.009 −0.308 2880

2.493 1.014 1.556 0.597 0.131 −0.652 −0.645 0.034 0.019 −0.264 2846

a Calculated values (B3LYP/aug-cc-pVTZ), r = interatomic distances, Ǻ ; Q = Wiberg bond orders; q = net atomic charges, e;̅ ω = harmonic frequencies, cm−1. bThe charge of the −C(O)−C−C(OH)− fragment.

our previous study4 we have noted strong correlations between several values, such as r(O−H), r(O···H), r(O···O), Q(O−H), ω(O−H), etc. Also we have concluded that the strength of the hydrogen bond does not depend on the electronegativities of the substituents.4 The data of Table 5 confirm our conclusions and demonstrate that the presence of one (in BA) or two (in DBM) phenyl groups does not change significantly the parameters of the H-bond. Furthermore, the overall net charge of the enol skeleton in BA and DBM are nearly equal to that in AcAc. Thus, no additional electron donating effect of the phenyl groups compared to that of the methyl groups occurs. Therefore, apparently (according to r(O−H) and r(O···O) values) strengthening of the intramolecular hydrogen bond in BA and DPM in comparison with that in AcAc can hardly be explained by the influence of phenyl groups through hyperconjugation with the enol ring.



EXPERIMENTAL SECTION The electron diffraction patterns and the mass spectra were recorded simultaneously using the techniques described previously.31,32 Conditions of the GED/MS experiment and the relative abundance of the characteristic ions of benzoylacetone are shown in Tables 6 and 7, respectively. The main ions in the mass spectra are the same with the literature data for BA.33 The temperature of the effusion cell 3433

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

tions, thermal energies, enthalpies, and Gibbs free energies for two enol forms (enol1 and enol2). This material is available free of charge via the Internet at http://pubs.acs.org.

Table 6. Conditions of GED Experiment nozzle-to-plate distance, mm fast electron beam, μA temperature of effusion cell, K accelerating voltage, kV (electron wavelength), Ǻ ionization voltage, V exposure time, s residual gas pressure, Torr

338 0.99 331(5) 71 (0.044590(44)) 50 80−90 1.8 × 10−6

598 0.52 331(5) 68 (0.045460(37)) 50 60−70 2.2 × 10−6



Corresponding Author

*E-mail:[email protected]. Notes

The authors declare no competing financial interest.



Table 7. Mass Spectral Data of the Vapor of Benzoylacetone m/e

ion

abundance, %

162 147 105 85 77 69 51 44 15

[M]+ [M − CH3]+ [C(O)C6H5]+ [M − C6H5]+ [C6H5]+ [M − CH3−C6H5−H]+ [C6H5−C2H2]+ [CH3C(O)]+ [CH3]+

51 39 32 18 50 100 25 50 23

ACKNOWLEDGMENTS This work was supported by the Deutsche Forschungsgemeinschaft, the Fonds der Chemischen Industrie) and by the Russian-German Cooperation Project (Russian Foundation for Basic Research Grant N 09-03-91341-NNIO_a and DFG OB 28/22-1)



REFERENCES

(1) Gomez-Garibay, F.; Calderon, J. S.; Quijano, L.; Tellez, O.; Soccoro-Olivares, M.; Rios, T. Phytochemistry 2001, 46, 1285−1287. (2) Wetz, F.; Routaboul, C.; Lavabre, D.; Garrigues, J. C.; RicoLatters, I.; Pernet, I.; Denis, A. Photochem. Photobiol. 2004, 80, 316. (3) Emsley, J. In Structure and bonding; Clarke, M. J., J. B. G., Ibers, J. A., Jorgensen, C. K., Mingos, D. M. P., Neilands, J. B., Reinen, D., Sadler, P. J., Weiss, R., Williams, R. J. P., Eds.; Complex Chemistry, Springer-Verlag: Berlin, 1984; Vol. 57, pp 147−199. (4) Belova, N. V.; Sliznev, V. V.; Oberhammer, H.; Girichev, G. V. J. Mol. Struct. 2010, 978, 282−293. (5) Allen, G.; Dwek, R. A. J. Chem. Soc. B. 1966, 2, 161−163. (6) Sardella, D. J.; Heinert, D. H.; Shapiro, B. L. J. Org. Chem. 1969, 34, 2817−2821. (7) Hay, R. W.; Williams, P. P. J. Chem. Soc. 1964, 2270−2272. (8) Bassetti, M.; Cerichelli, G.; Floris, B. Tetrahedron 1988, 44, 2997−3004. (9) Matsuzawa, H.; Nakagaki, T.; Iwahashi, M. J. Oleo Sci. 2007, 56, 653−658. (10) Burdett, J. L.; Rogers, M. T. J. Am. Chem. Soc. 1964, 86, 2105− 2109. (11) Shapet’ko, N. N.; Berestova, S. S.; Lukovkin, G. M.; Bogachev, Y. S. Org. Magn. Reson. 1975, 7, 237−239. (12) Morita, H.; Nakanishi, H. Bull. Chem. Soc. Jpn. 1981, 54, 378− 386. (13) Bunting, J. W.; Kanter, J. P.; Nelander, R.; Wu, Z. Can. J. Chem. 1995, 73, 1305−1311. (14) Chan, S. I.; Lin, L.; Clutter, D.; Dea, P. Proc. Natl. Acad. Sci. U. S. A. 1970, 65, 816. (15) Gorodetsky, M.; Luz, Z.; Mazur, Y. J. Am. Chem. Soc. 1967, 89, 1183. (16) Borisov, E. V.; Skorodumov, E. V.; Pachevskaya, V. M.; Hansen, P. E. Magn. Reson. Chem. 2005, 43, 992−998. (17) Gilli, P.; Bertolasi, V.; Pretto, L.; Ferretti, V.; Gilli, G. J. Am. Chem. Soc. 2004, 126, 3845−3855. (18) Madsen, G. K. H.; Iversen, B. B.; Larsen, F. K.; Kapon, M.; Reisner, M.; Herbstein, F. H. J. Am. Chem. Soc. 1998, 120, 10040− 10045. (19) Herbstein, F. H.; Iversen, B. B.; Kapon, M.; Larsen, F. K.; Madsen, G. K. H.; Reisner, M. Acta Crystallogr. 1999, B55, 767−787. (20) Semmingsen, D. Acta Chem. Scand. 1972, 26, 143−154. (21) Tayyari, S. F.; Emampour, J. S.; Vakili, M.; Nekoei, A. R.; Eshghi, H.; Salemi, S.; Hassanpour, M. J. Mol. Struct. 2006, 794. (22) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; Vreven, J. T.; Kudin, K. N.; Burant, J. C.;et al. Gaussian03; Gaussian Inc.: Pittsburgh, PA, 2003. (23) Sipachev, V. A. J. Mol. Struct. 2001, 567−568, 67−72. (24) Sipachev, V. A. J. Mol. Struct. 1985, 121, 143−151.

was measured by a W/Re-5/20 thermocouple calibrated by the melting points of Sn and Al. The wavelength of electrons was determined from diffraction patterns of polycrystalline ZnO. Optical densities were measured by a computer controlled MD100 (Carl Zeiss, Jena) microdensitometer.34 The molecular intensities sM(s) were obtained in the s-ranges 3.6−26.6 and 1.5−14.2 Ǻ −1 for the short and long nozzle-to-plate distance, respectively (s = (4π/λ) sin θ/2, λ is electron wavelength and θ is scattering angle). Experimental and theoretical intensities sM(s) are compared in Figure 4.

Figure 4. Experimental (dots) and calculated (solid lines) modified molecular intensity curves and residuals (experimental − theoretical) for enol2 at two nozzle-to-plate distances (L1 = 598 mm, L2 = 338 mm).



AUTHOR INFORMATION

ASSOCIATED CONTENT

S Supporting Information *

Interatomic distances, vibrational amplitudes, and vibrational corrections for the enol forms (excluding nonbonded distances involving hydrogen), the results of B3LYP/cc-pVTZ calculations: Cartesian coordinates, frequencies, zero-point correc3434

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435

The Journal of Physical Chemistry A

Article

(25) Sipachev, V. A. In Advanced in Molecular Structure Research; Hargittai, I., Hargittai, M., Eds.; JAI Press: New York, 1999; Vol. 5, p 263−311. (26) Glendening, E. D.; Badenhoop, J., K. ; Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Weinhold, F. Theoretical Chemistry Institute, University of Wisconsin: Madison, WI, 2004; http://www.chem.wisc.edu/∼nbo5. (27) Granovsky, A. A. www http://classic.chem.msu.su/gran/firefly/ index.html. (28) Belova, N. V.; Sliznev, V. V.; Oberhammer, H.; Girichev, G. V. J. Mol. Struct. 2009, in press. (29) Grens, E.; Grinvalde, A.; Stradins, J. Spectrochim. Acta 1975, 31A, 555−564. (30) Tayyari, F.; Rahemi, H.; Nekoei, A. R.; Zahedi-Tabrizi, M.; Wang, Y. A. Spectrochim. Acta A 2007, 66, 394−404. (31) Girichev, G. V.; Shlykov, S. A.; Revichev, Y. F. Prib. Tekh. Eksp. 1986, 4, 167. (32) Girichev, G. V.; Utkin, A. N.; Revichev, Y. F. Prib. Tekh. Eksp. 1984, 2, 187. (33) Bowie, J. H.; Williams, D. H.; Lawesson, S.-O.; Schroll, G. J. Org. Chem. 1966, 31, 1384−1390. (34) Girichev, E. G.; Zakharov, A. V.; Girichev, G. V.; Bazanov, M. I. Izv. Vyssh. Uchebn. Zaved., Tekhnol. Tekst. Prom-sti. 2000, 2, 142−146.

3435

dx.doi.org/10.1021/jp2118967 | J. Phys. Chem. A 2012, 116, 3428−3435